How to Write a Research Protocol for Veterinary Studies

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• A robust veterinary research protocol is a prospective blueprint that explicitly defines the study population (species, breed, age, sex, health status, husbandry), intervention(s) (drug, surgical technique, management change), comparator, and measurable, biologically meaningful outcome(s) using the PICO framework to ensure testability and prevent overly broad research questions.
• Study design selection (prospective vs. retrospective, experimental vs. observational) must logically follow the research question to support causal inference or accurately capture existing data, with prospective designs enabling standardization of measurements and blinding, while retrospective designs are limited by original record quality.
• Pre-specification of the primary outcome, sample size calculation (based on effect size, variability, significance level, and power), eligibility criteria (inclusion/exclusion), blinding procedures, and statistical analysis plan (including handling of missing data and humane endpoints) is critical before data collection to ensure study integrity and prevent post-hoc manipulation.
• Animal welfare is paramount, requiring detailed protocols for housing, husbandry, minimization of pain and distress, and clearly defined, observable humane endpoints that trigger early withdrawal or euthanasia, alongside adherence to relevant international (e.g., WOAH) and institutional (e.g., IACUC) regulatory frameworks.
• Protocol implementation necessitates translation into Standard Operating Procedures (SOPs) for consistent execution, with clear mechanisms for monitoring protocol deviations, managing data quality through source documentation and version control, and anticipating reporting requirements (e.g., ARRIVE, CONSORT, STROBE, REFLECT) from the outset.
• Recognized complications like endpoint drift, recruitment failure, measurement error, and missing data necessitate proactive monitoring, transparent reporting, and formal amendment processes for any changes to the protocol, ensuring the study's scientific validity and reproducibility.

A research protocol is the prospective blueprint that converts a clinical question into a reproducible investigation. It defines the study population, the interventions or exposures under examination, the outcome measures, the statistical analysis plan, and the procedures for safeguarding animal welfare. For veterinary researchers, the protocol serves three distinct functions: it forces explicit decisions before data collection begins, it provides the document against which ethics and regulatory review is judged, and it becomes the reference point for detecting protocol deviations during execution.

This article serves veterinary researchers who are designing primary studies across species, including companion animals, production animals, and laboratory species. It addresses the procedural steps of protocol construction, from framing the research question to specifying the analysis plan. The scope excludes grant writing, although many of the same elements will appear in funding applications. The emphasis throughout is on decisions that must be made before the first animal is enrolled, because the quality of those decisions determines whether the study can answer its question at all.

A well-constructed protocol also anticipates reporting. The ARRIVE guidelines, published by the NC3Rs, specify the minimum information required for transparent and reproducible animal research publications, and the EQUATOR Network maintains a library of reporting standards including CONSORT, PRISMA, STROBE, and REFLECT that apply to specific study designs. Writing the protocol with these reporting requirements in mind from the outset prevents the common failure of discovering, after data collection, that a critical variable was never recorded.

At a Glance

The Scientific Foundation of Protocol Design

The protocol rests on a single logical structure: the research question determines the design, the design determines the measurements, and the measurements determine the analysis. Disrupting any link in that chain compromises the entire study. The most common disruption is the reverse ordering, where a researcher chooses a convenient design or an available assay and then attempts to fit a question to it.

Framing the Research Question

The PICO framework, population, intervention, comparator, and outcome, provides a disciplined starting point. The population must be defined by species, breed, age, sex, health status, and husbandry or management system. The intervention may be a drug, a surgical technique, a management change, or an exposure. The comparator may be a placebo, a sham procedure, a standard treatment, or no intervention. The outcome must be measurable, biologically meaningful, and clinically relevant.

A question that cannot be expressed in PICO terms is usually not yet a research question. For example, "Does drug X work in dogs with osteoarthritis?" becomes testable when reframed as "In client-owned dogs with radiographically confirmed hip osteoarthritis, does drug X at the label dose improve owner-assessed mobility scores at 28 days compared with placebo?" The reframed version exposes every decision that must be made and every assumption that must be defended.

Selecting the Study Design

The design must match the question. Experimental designs, in which the investigator assigns the intervention, support causal inference but require careful control of confounding and carry the burden of welfare justification. Observational designs, in which the investigator records exposures and outcomes as they occur, cannot establish causation with the same certainty but may be the only ethical or feasible option for harmful exposures, rare outcomes, or production-scale questions.

The choice between prospective and retrospective designs deserves particular attention. Prospective designs allow the investigator to standardize measurements, ensure blinding, and capture all relevant covariates. Retrospective designs are faster and cheaper but inherit the limitations of the original records, including missing data, inconsistent measurement, and undocumented confounding. A retrospective design should be chosen only when the research question can be answered from data that were recorded with sufficient quality and completeness.

Defining Outcomes and Endpoints

The primary outcome must be specified before data collection. It should be the single measure that, if changed, would alter clinical practice. Secondary outcomes answer subsidiary questions but cannot substitute for the primary outcome in the sample size calculation or the main analysis.

The distinction between a surrogate endpoint and a clinical endpoint is critical. A surrogate, such as a biomarker or an imaging finding, may be easier to measure but may not reflect what patients or owners actually experience. The Acute Dialysis Quality Initiative group, in its consensus on acute renal failure, noted that more than 30 different definitions had been used in the literature, creating confusion and making comparisons difficult. That observation applies broadly: the same condition can be defined in many ways, and the protocol must state which definition is used and why.

The Protocol Document Structure

A complete protocol follows a standard structure that mirrors the logic of the study. The sections below describe the content expected in each part of the document.

Title, Abstract, and Background

The title should state the population, intervention, comparator, and outcome in enough detail to be informative without being unwieldy. The abstract summarizes the protocol in approximately 300 words, covering the question, design, population, outcomes, and analysis. The background section justifies the study by summarizing what is known, identifying the gap in knowledge, and explaining why the gap matters. It should cite the relevant literature but should not attempt a systematic review.

Objectives and Hypotheses

The objectives restate the research question as declarative aims. The hypotheses must be stated in falsifiable form. A null hypothesis and an alternative hypothesis should be explicit, because they determine the statistical tests and the interpretation of results. For non-inferiority or equivalence designs, the margin of clinical equivalence must be defined in advance and justified from clinical reasoning or published evidence.

Study Population and Eligibility

The source population is the group to which the results will be generalized. The study population is the group from which participants are drawn. The eligibility criteria define who can enter the study and who cannot. Inclusion criteria should be broad enough to support generalization but narrow enough to reduce heterogeneity. Exclusion criteria should remove individuals for whom the intervention is contraindicated, who cannot comply with the protocol, or who would introduce unacceptable confounding.

For production animal studies, the unit of interest may be the animal, the pen, the herd, or the flock. The protocol must state the unit of allocation and the unit of analysis, because these determine the statistical methods and the interpretation of results. Cluster-randomised designs, in which groups instead of individuals are allocated to treatment, require larger sample sizes and specialised analysis methods.

Interventions and Procedures

The intervention section must describe the treatment, the dose, the route, the frequency, and the duration in sufficient detail that another investigator could replicate the study. For surgical interventions, the protocol should specify the technique, the anesthetic protocol, the perioperative care, and the criteria for postoperative intervention. For management interventions, the protocol should specify the husbandry, feeding, housing, and monitoring procedures.

The comparator must be described with equal precision. A placebo should be matched for appearance, route, and schedule. A sham procedure should replicate the non-therapeutic aspects of the intervention. The protocol should also describe permitted concomitant treatments and the circumstances under which rescue therapy may be administered.

Outcome Measurement and Follow-up

Each outcome requires a definition, a measurement method, a measurement schedule, and a specification of who performs the measurement. The protocol should state whether outcome assessors are blinded to treatment allocation and how blinding is maintained. For subjective outcomes, such as lameness scores or owner questionnaires, the protocol should specify the scale, the training of assessors, and the procedures for ensuring consistency.

The follow-up schedule must be specified in calendar time or in relation to the intervention. The protocol should define the duration of follow-up, the timing of each assessment, and the procedures for handling losses to follow-up. The distinction between the per-protocol population and the intention-to-treat population should be stated, along with the primary analysis population.

Sample Size and Statistical Analysis

The sample size calculation requires four inputs: the expected effect size, the variability of the outcome, the significance level, and the statistical power. The protocol must state each input, the source of the estimate, and the formula or software used. The calculation should be presented transparently so that reviewers can reproduce it.

The statistical analysis section must pre-specify the primary analysis, the secondary analyzes, the handling of missing data, and the procedures for interim analysis. The analysis should be described in enough detail that a statistician could implement it without further instruction. The protocol should also state the software to be used and the threshold for statistical significance.

Animal Welfare and Ethical Considerations

The welfare section must describe the housing, husbandry, and environmental conditions, the procedures for minimizing pain, distress, and suffering, and the humane endpoints that trigger early withdrawal from the study. The humane endpoints should be defined in terms of observable clinical signs, physiological parameters, or behavioral changes, and they should be specified before the study begins. The protocol should also describe the monitoring schedule, the personnel responsible for welfare assessment, and the procedures for veterinary intervention.

The protocol must acknowledge the regulatory framework under which the study will be conducted. The World Organization for Animal Health publishes international standards for animal health and welfare in its Terrestrial Animal Health Code, and national and regional authorities impose additional requirements. The protocol should identify the relevant approvals required, including institutional animal care and use committee review, and should describe how the study will comply with applicable standards.

Data Management and Quality Assurance

The data management section describes how data will be recorded, stored, backed up, and protected. It should specify the data collection forms, the database structure, the procedures for data entry and verification, and the arrangements for data ownership and sharing. The quality assurance section describes the procedures for monitoring protocol compliance, detecting errors, and handling deviations.

Dissemination and Reporting

The protocol should state the plans for disseminating the results, including publication in the peer-reviewed literature, presentation at conferences, and communication to stakeholders. The reporting plan should identify the relevant reporting guideline from the EQUATOR Network collection, such as CONSORT for randomised trials, PRISMA for systematic reviews, STROBE for observational studies, or REFLECT for livestock studies, and should commit to following it.

Protocol Implementation and Monitoring

A research protocol is a living document. Once approved, it governs every subsequent decision, from animal acquisition to data archiving. The practical value of the protocol depends on how completely it anticipates the conditions of execution.

Operationalising the Protocol

The first implementation step is translating the protocol into standard operating procedures (SOPs). Each SOP should specify the personnel responsible, the equipment required, and the sequence of actions for a single procedure. For example, a blood sampling SOP must state the collection site, the order of tubes, the maximum allowable collection volume, and the handling time before centrifugation. The SOPs become the training documents for study staff and the audit trail for quality assurance.

The protocol should identify which deviations require immediate notification of the principal investigator and which can be recorded and reviewed at scheduled monitoring visits. Define the threshold for a major deviation, such as an animal receiving the wrong treatment or a sample being lost, and the procedure for documenting it. Minor deviations, such as a blood draw occurring 30 minutes outside the window, should still be recorded with the reason.

Monitoring Parameters and Safety Triggers

The protocol must specify the parameters that will be monitored during the study and the thresholds that trigger intervention. These parameters fall into three categories: those that assess animal welfare, those that assess treatment efficacy, and those that assess data quality.

For welfare monitoring, the protocol should define a scoring system that combines clinical observations. A typical system assigns scores for demeanour, appetite, respiratory effort, and mobility. The protocol must state the cumulative score at which an animal is removed from the study, receives rescue analgesia, or is euthanised. These humane endpoints should be defined before the study begins, not in response to an adverse event.

The table below presents a framework for welfare monitoring parameters and their decision thresholds.

The thresholds in this table are illustrative. The correct values depend on the species, the model, and the expected course of the disease being studied. For a study of a rapidly progressive condition, the monitoring interval may need to be shortened to every 4 hours. For a production animal study, the thresholds may be adjusted to reflect the economic and welfare realities of the setting.

Equipment and Consumable Selection

The protocol should specify the equipment and consumables required for each procedure, including the manufacturer and catalogue number where relevant. This level of detail serves two purposes. It ensures that all sites and personnel use identical materials, and it allows the study to be reproduced by other groups.

For assays and laboratory procedures, the protocol should state the analytical method, the reference range for the laboratory, and the quality control procedures. If the study uses a novel assay, the protocol should include the validation data or reference the publication that established the method. For example, a study using simulated gastrointestinal digestion should specify the exact enzyme concentrations, pH values, and incubation times, following the standardized method developed by the international consensus group Minekus et al., 2014.

Species and Setting Adaptations

The correct monitoring parameters and equipment choices differ substantially across species and production systems. A laboratory rodent study allows continuous observation and precise environmental control. A field study in cattle requires protocols that work within the constraints of the farm, such as handling facilities, milking schedules, and the availability of trained personnel.

For companion animal studies, client compliance is a major factor. The protocol should specify how owners will be trained to administer treatments and record observations, and how the study team will verify compliance at each visit. For production animal studies, the protocol must align with the standards of the relevant international body, such as the World Organization for Animal Health terrestrial animal health code, which addresses disease surveillance, welfare, and trade-related considerations.

Data Collection Forms and Source Documentation

The protocol should include or reference the data collection forms that will be used. These forms are the source documents for the study. They should be designed before the study begins and piloted on a small number of animals to identify ambiguous fields or missing categories.

Each form should include the animal identification number, the date and time, the initials of the person completing the form, and the protocol visit number. The form should use closed fields wherever possible, such as checkboxes or numeric entry boxes, to reduce transcription errors. Free-text fields should be reserved for unexpected observations.

The protocol should state the policy for correcting errors in source documents. The standard approach is to draw a single line through the error, write the correction, initial and date the change, and leave the original entry legible. Corrections should never be made by erasing or covering the original entry.

Protocol Amendments and Version Control

The protocol will require amendments during the study. The amendment process should be specified in the protocol itself. Each amendment must be numbered, dated, and approved by the same bodies that approved the original protocol, such as the institutional animal care and use committee and any funding agency.

The protocol should state which changes require an amendment and which can be implemented as administrative updates. Changes to the study population, the intervention, the outcome measures, or the sample size always require a formal amendment. Changes to contact details or the names of study personnel can be handled as administrative updates.

Version control is critical. The protocol should include a version history table that lists the version number, the date, the author, and a summary of changes. All study personnel must be trained on the current version, and obsolete versions must be removed from circulation.

The Completeness Checklist

A protocol is complete when it can answer the following questions without reference to external documents. Use this checklist during the final review.

• Does the title identify the study design, the population, and the intervention?
• Does the background cite the key literature and state the knowledge gap?
• Are the objectives stated as measurable hypotheses?
• Is the primary outcome defined with a specific measurement method and time point?
• Are the eligibility criteria explicit, including the justification for exclusions?
• Is the sample size calculation reproducible from the stated assumptions?
• Are the randomisation and allocation concealment methods described?
• Are the interventions specified with dose, route, frequency, and duration?
• Are the welfare monitoring parameters and humane endpoints defined?
• Are the data management procedures, including backup and access control, described?
• Are the statistical methods specified for the primary and secondary outcomes?
• Is the dissemination plan stated, including the target journals and the data sharing approach?

The reporting guidelines maintained by the EQUATOR Network provide a useful cross-check. The ARRIVE guidelines, which specify the minimum information required for transparent animal research publications, can be used to verify that the protocol contains the elements that will be needed for the final manuscript ARRIVE guidelines 2.0. A protocol that satisfies the reporting requirements at the design stage will produce a manuscript that meets them at the publication stage.

Recognized Complications and Failure Modes

Protocols fail in predictable patterns. The most common is endpoint drift, where the research team modifies outcome definitions after data collection begins because the original endpoints prove difficult to measure. This compromises the statistical integrity of the study. Detect it early by requiring that any endpoint change be logged as a formal amendment with justification, instead of an informal note in a laboratory book.

Recruitment failure occurs when eligibility criteria are too restrictive for the available caseload. A protocol that requires 120 dogs with a rare comorbidity will stall. Monitor recruitment rates monthly against projections. If enrollment falls below 80% of target for two consecutive months, the eligibility criteria or the recruitment strategy requires revision.

Measurement error arises from inconsistent technique between operators. This is particularly problematic in studies using ultrasound, endoscopic scoring, or behavioral assessment. The discriminating check is a reproducibility exercise: have two operators score the same ten animals or samples independently and calculate agreement statistics before the main study begins.

Missing data accumulates silently. Animals withdraw, samples hemolyse, and recording sheets go missing. The failure mode is that analysis proceeds without accounting for these losses. Detect it by auditing data completeness at each scheduled monitoring visit. The ARRIVE 2.0 reporting guidelines require transparency about animal losses and exclusions, which means the protocol must specify how missing data will be handled before the study starts.

Common Errors and Corrective Actions

Less experienced researchers often confuse statistical significance with clinical importance. A study may detect a statistically significant difference that is too small to matter in practice. Correct this by defining the minimum clinically important difference during the sample size calculation and interpreting results against that threshold.

Another frequent error is inadequate blinding. Investigators who know the treatment allocation may unconsciously influence outcome assessment, particularly for subjective measures such as pain scores or radiographic grading. The corrective action is to separate the team members who administer treatments from those who assess outcomes, and to document who was blinded to what.

Protocols frequently under-specify the study population. Terms such as "healthy adult dogs" leave too much to interpretation. Specify breed, age range, body condition score, vaccination status, and exclusion criteria such as concurrent medication. The MSD Veterinary Manual provides species-specific reference ranges that can anchor these definitions.

A third error is overpromising in the timeline. Veterinary studies depend on caseload, seasonal disease patterns, and breeding cycles. Build in a 20% buffer for unexpected delays and specify what happens if the study cannot complete within the planned period.

Limitations of Current Evidence

The veterinary evidence base has structural gaps. Many interventions are supported by observational studies instead of randomised trials, and the EQUATOR Network reporting guidelines catalogue the specific reporting standards that apply to each study type. A protocol for a randomised trial should follow CONSORT, while an observational study should follow STROBE. The REFLECT statement covers livestock trials.

Expert opinion still differs on several points. The choice of primary endpoint in pain studies remains contested, with some groups favouring composite measures and others preferring single validated scales. Similarly, there is no universal consensus on the most appropriate animal models for translational research, as the Acute Dialysis Quality Initiative consensus on acute renal failure demonstrated for one specialty. That group identified more than 30 different definitions of a single condition, which made cross-study comparison impossible.

In vitro methods carry their own uncertainties. Standardized digestion models have improved reproducibility, but the international consensus on static in vitro digestion acknowledges that results vary with enzyme source, pH, and ionic strength. Extrapolating in vitro findings to whole animals requires caution.

Escalation and Referral

Certain situations warrant escalation beyond the research team. If an adverse event occurs that is unexpected, fatal, or suggests a previously unrecognised risk, the institutional animal ethics committee must be notified promptly. The protocol should name the responsible person for this notification and specify the timeframe.

Regulatory reporting obligations vary by jurisdiction and by product type. Studies involving investigational veterinary products, feed additives, or biologics may trigger reporting requirements to national regulatory authorities. The WOAH terrestrial animal health standards address disease surveillance and trade-related reporting that may apply to studies involving notifiable diseases.

Laboratory involvement becomes necessary when results fall outside expected ranges or when sample quality is questionable. The protocol should specify which laboratory tests require confirmation by a reference laboratory and what actions follow discrepant results.

Specialist consultation is warranted when the study population includes animals with complex comorbidities, when a novel intervention carries uncertain risk, or when the research question extends beyond the team's expertise. The AVMA practice resources can help identify relevant specialty organizations and referral pathways.

Troubleshooting Table

Frequently Asked Questions

How do I adjust my protocol when the ideal equipment or consumables are unavailable?

Substitute only after documenting that the alternative does not compromise validity. For laboratory procedures, validate the substitute against the reference method in a pilot run and report the comparison. For field studies, record lot numbers, supplier, and storage conditions for every consumable batch. When a standardized method exists, such as the international consensus static digestion protocol, state explicitly where your materials deviate from that method and justify the deviation. If the substitution affects measurement precision or detection limits, recalculate sample size or acknowledge the reduced power in the limitations section. Escalate unresolved substitutions to the study sponsor or ethics committee before commencing data collection.

What should I do when recruitment is slower than projected?

Revisit the eligibility criteria and recruitment strategy before extending the timeline. Calculate the recruitment rate per site per month and compare it with the rate assumed in the sample size calculation. If the shortfall exceeds 20 percent, consult the statistician to assess conditional power. Consider adding recruitment sites, broadening referral networks, or relaxing narrow inclusion criteria if the scientific question tolerates it. Document every change as a protocol amendment with version control. If the target remains unattainable, pre-specify the stopping rule and report the study as underpowered instead of presenting exploratory results as confirmatory. Transparent reporting standards require this level of disclosure.

How do I handle a serious adverse event during the study?

Follow the safety trigger thresholds written into the protocol before the study began. Pause enrollment for that cohort, notify the institutional animal care and use committee within the time frame specified in your approval, and initiate the escalation pathway. Separate the immediate clinical response to the animal from the study integrity response. The clinical team treats the patient, the investigator team reviews whether the event meets the protocol definition of an expected or unexpected adverse event. If the event is unexpected, file an amendment and reassess the risk-benefit balance. Do not unblind the treatment allocation unless the treating clinician determines that knowledge of the allocation changes clinical management.

How do I budget for a veterinary research protocol?

Build the budget from the protocol procedures, not from a round figure. Itemise animal acquisition and housing, feed, consumables, diagnostics, personnel time, equipment access, and data management. Add a contingency line of 10 to 15 percent for consumable price changes and repeat assays. For multi-site studies, include site initiation and monitoring visits. If the study uses specialised imaging or laboratory services, obtain written quotes before finalising the budget. The AVMA practice resources provide guidance on professional fee structures and practice-based research costs. Separate the budget from the grant application narrative, since the protocol must stand alone as a costed document for ethics review and institutional approval.

How do I adapt a protocol designed for one species to another species?

Re-derive every parameter instead of scaling by body weight alone. Metabolic rate, drug disposition, and susceptibility to specific adverse effects differ across species, and the MSD Veterinary Manual provides species-specific pharmacology and clinical guidance that should inform these adaptations. Re-examine the eligibility criteria, because age, sex, and reproductive status have different relevance in each species. Validate all measurement instruments for the new species, including reference intervals for clinical pathology. Recalculate the sample size if the expected effect size or variance changes. For food-producing animals, consult the WOAH terrestrial animal health standards for disease surveillance and trade implications before finalising the protocol.

How do I explain protocol deviations to a supervisor or sponsor?

Report deviations immediately and in writing, not at the next scheduled meeting. State what happened, when it happened, which protocol section was violated, and what data were affected. Classify the deviation as minor or major according to the criteria in the protocol. Minor deviations, such as a blood sample drawn 30 minutes late, are documented in the study file. Major deviations, such as an unapproved eligibility waiver, require a formal corrective action plan. Use the deviation as a trigger to review whether the protocol itself was ambiguous or impractical. If the protocol caused the error, amend it. If the staff caused the error, retrain them. The EQUATOR Network reporting guidelines describe how deviations should be reported in the final manuscript.

Related Clinical & Scientific Guides

• Conducting Systematic Reviews of Veterinary Diagnostic Test Accuracy
• Bias in Veterinary Research: Types, Sources, and Mitigation
• Cluster Randomized Trials in Veterinary Research: Design and Analysis

References and Further Reading

• Acute renal failure - definition, outcome measures, animal models, fluid therapy and information technology needs: the Second International Consensus Conference of the Acute Dialysis Quality Initiative (ADQI) Group.. 2004.
• Quantum dots for live cells, in vivo imaging, and diagnostics.. 2005.
• Randomized trial of estrogen plus progestin for secondary prevention of coronary heart disease in postmenopausal women. Heart and Estrogen/progestin Replacement Study (HERS) Research Group.. 1998.
• A standardized static in vitro digestion method suitable for food - an international consensus.. 2014.
• ARRIVE Guidelines 2.0 for Reporting Animal Research. PLOS Biology, 2020.
• EQUATOR Network Reporting Guidelines. EQUATOR Network.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.
• American Veterinary Medical Association Practice Resources. American Veterinary Medical Association.
• WOAH Terrestrial Animal Health Code. WOAH.

Related Articles

• Designing Questionnaire Studies for Veterinary Research
• Cohort Studies in Veterinary Research: Design and Interpretation
• Using Mixed Methods in Veterinary Research
• Conducting Pharmacovigilance Studies in Veterinary Medicine
• Conducting Qualitative Research in Veterinary Settings

This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.